Bis(trifluoromethyl)dithiophosphinic acid and related derivatives

Ronald W. Mitchell , Max Lustig , Frederick A. Hartman , John K. Ruff , James A. Merritt. Journal of the American Chemical Society 1968 90 (23), 6329-...
0 downloads 0 Views 617KB Size
2011

ing region7317 and is similar t o the spectrum of diisopropoxyborane, but attempts to isolate this compound from this reaction mixture by conventional high-vacuum techniques, to date, have been unsuccessful. This may be due t o the subsequent reaction of the alkoxyborane with morpholine, a by-product of the ketone reduction. Numerous reactions of amines with borate esters have been reported,1s-22 and we have found that diiso(17) L. J. Bellamy, W. Gerrard, M. F. Lappert, and R. L. Williams, J . Chem. Soc., 2412 (1958), and references therein. (18) H. K. Zimmerman in “Boron-Nitrogen Chemistry,” Advances in Chemistry Series, No. 42, American Chemical Society, Washington, D. C., 1964, Chapter 3, and references therein. (19) E. J. Mezey, P. R. Giradot, and W. E.Bissinger, ref 18, Chapter 19.

propoxyborane does, in fact, react with morpholine in acetone solution to produce a hygroscopic white solid and slowly evolve hydrogen. This product is being investigated further. Acknowledgment. We wish to acknowledge the T.C.U. Research Foundation for support of this work which included a research fellowship for S. S. W. Also we wish to thank a referee whose discussion led to a study of the rate of the morpholine-borane-2-propanol reaction. (20) K. Niedenzu and J. W. Dawson, “Boron-Nitrogen Compounds,” Academic PressInc., New York, N. Y., 1965. (21) S. V. Urs and E.S.Gould, J . Am. Chem. Soc., 74,2948 (1952). (22) R. J. BrothertonandH. Steinberg,J. Org. Chem.,26,4632(1961).

Bis ( trifluoromethyl)dithiophosphinic Acid and Related Derivatives’ Keith Gosling and Anton B. Burg

Contribution from the Department of Chemistry, University of Southern California, Los Angeles, California 90007. Received September 11, 1967 Abstract: The readily volatile new dithiophosphinic acid (CF3)2PSzH(mp 14”; bp estd 105” ; monomeric, unlike the analogous oxygen compound) has been made by the HzS cleavage of the nonvolatile [(CF3)zPS]zS(presumed), obtained by heating [(CF&P],S with sulfur. Another process, in which (CF&P(S)I is attacked by HzS, depends upon a novel reduction process, liberating iodine. The spontaneous loss of iodine from (CF3),P(S)I was pushed forward by mercury, giving a 66z yield of [(CF3)2P]2S;probably the reaction involves an anti-Arbuzov type of rearrangement. Other new derivatives of the dithiophosphinic acid include the chloride, the methyl ester, and two amides. Also reported are the thiophosphonic dichloride CF3P(S)CIzand the corresponding bis(dimethy1amide). All of these have been fully characterized in regard to melting point, volatility, and infrared spectra, with P=S stretching and bending frequencies demonstrating a theoretically interesting trend of bond order.

T

he (CF3)2Pphosphines and phosphinic acid derivatives differ enough from the analogous R2P compounds (R = ordinary alkyl or aryl) to suggest that the chemistry of the compounds of type (CF3)2P(S)X (X = SH, SR, C1, I, or an amido group) may not be fully predictable. Accordingly, we have made and studied such thiophosphinic compounds, as well as two examples of the thiophosphonic type CFsP(S)X2. The new dithiophosphinic acid (CF3)2PS2Hwas obtained by two methods: addition of sulfur to [(CF3)2P ] z S at ~ ~150” ~ to form a nonvolatile liquid assumed to be [(CF&PSI2S, which is cleaved by hydrogen sulfide at 100”; or by photochemical formation of the iodide (CF&P(S)I from [(CF3),PI2S and Iz, followed by an interesting reaction with H2S, whereby half of the phosphinic material is reduced to the phosphinous state. S 2(CFa)zP-I

S (CFs)zP-SH

S

+ 2HzS +2HI + 2(CF&P-SH + 2HI +(CFs)zPSH + HzS + Iz

Apparently the reduction is the driving force for this (1) Supported by Grants GP-199, GP-3812, and GP-6751X from the National Science Foundation, which assisted also toward providing the instruments required for some parts of this research. (2) R. G. Cave11 and H. J. Emeldus, J . Chem. Soc., 5825 (1964). (3) A. B. Burg and K. Gosling, J . Am. Chem. SOC.,87, 2113 (1965).

result, for the analogous reaction of (CF&P(S)Cl does not occur. Although obtainable almost quantitatively as a colorless liquid, the iodide (CF,),P(S)I easily loses iodine, with interesting chemical results. Using mercury to remove the iodine, one might expect to obtain the diphosphine disulfide, but this would have good reason to undergo an anti-Arbuzov type of rearrangement4

s s

S

(CFdJ’-P(CFdz

+(CFdzP-S-P(CF3)z

after which the resulting P-S-P compound would be expected to exchange groups in at least two ways S 2(CFM’-S--P(CFdz

s s

+(CF3)zP--P(CFdz

+

(CFdzP-S-S--P(CFs)z S 2(CFdzP--S--P(CFdz

S

+(CFdzP-P(CFdz

+

s

s

(CF&P-S-P(CFa)z

Then the anti-Arbuzov rearrangement of the diphosphine monosulfide would produce (CF3)2PSP(CF3)2, which would come also from the action of mercury upon the P-S-S-P compound, and the diphosphine disulfide would go through the whole process again. If the two group exchanges should occur in equal (4) J. E. Griffiths and A. B. Burg, Proc. Chem. Soc., 12 (1961).

Gosling, Burg / Bis(tri8uoromethyl)dithiophosphinic Acid

2012

tube (resealed), for 4 hr at 100"; then the volatile product was almost exclusively (CF&PS2H. The yield was not measured accurately, but certainly exceeded 85 %. The two-stage process seemed necessary, for the initial presence of HzS in a single-stage run would have meant the conversion of much of the (CF,),PSP(CF& to (CF3)2PSH, more rapidly than S could be attached to the (CF3)2P unit; and experiments on the direct addition of sulfur to (CF&PSH always gave impure (CF&PS2H, not purifiable by high-vacuum distillation methods. Also necessary was the use of excess sulfur for conversion of (CF3)2PSP(CF3)z to the supposed (CF&PS,P(CF&, for an experiment using only 1.993 mg-atoms of sulfur per 1.833 mmoles of (CF&PSP(CF& (26 hr, 165") gave very little of the nonvolatile liquid, but much of an unstable liquid which may have been (CF3)?PS2P(CF3)2. Its vapor tension at 0" was near 1 mm, and it deposited a nonvolatile white solid (combustion of which in Oz gave no SOz) wherever it was condensed in the high-vacuum line. It reacted with H,S during 1 hr at 50" to form (CF3)zPSH, (CF&PSP(CF&, and a (CF&PSZH fraction which could not be purified by distillation. The Thiophosphinic Acid Chloride. A sealed tube with 11.6 mmoles of (CF&PSP(CFs)2 and 24.6 mmoles of Cl? was warmed from -196 to -50" during 7 hr. The unused Cl? was removed through a -78" trap under high vacuum; then the major part of the product (CF&P(S)Cl was similarly distilled through a trap at -60". The yields were 10.71 mmoles of (CF3),P(S)C1 (9373 and 11.3 mmoles of (CF&PCla (98%). The separation of these was done by means of a micro-size reflux column at - 25" (avoiding solid formation), with the reflux ratio controlled by a stopcock leading into the high-vacuum system. Alternatively, 0.81 mmole of (CF&PC13 and 1.22 mmoles of dry Ag2S were heated in a very small sealed tube (keeping a liquid phase) for 4.5 days at 100". The yield of (CF3)2P(S)C1 was 0.78 mmole (96%). Further heating with AgzS (5 days at 140") had no effect: there was no formation of either a silver thiophosphinate or a sulfur analog of the phosphinic anhydride. For the third synthesis, 3.89 mmoles of (CF&PCI and the equivalent amount of powdered sulfur ("flowers"), in a sealed tube with 0.344 mmole of freshly sublimed AlzC16,showed no reaction at 25" but a 3-min heating at 100" initiated a process which very rapidly consumed the sulfur, The yield of (CF&P(S)Cl was 3.79 mmoles, or 97.5%. Both this and the product of the silver sulfide method were purified by the micro-column under high vacuum with the reflux at -60". Syntheses and Characterizations This acid chloride proved to be inert to HSP(CF&, and an attempt to cause the formation of (CF&PThe Dithiophosphinic Acid. For the synthesis of pure (CF&PS2H, a sealed tube containing (CF3)2PSP(CF3)2 (S)SP(CF& by using trimethylamine to remove HC1 led only to the conversion of the HSP(CF3)z to (CF:)zwith an excess over 2 g-atoms of sulfur per mole was heated 46 hr at 160"; lower temperatures were inefPSP(CF3)?, with (CH&N. H2S as the probable byfective. The yield of a nonvolatile liquid, presumed to product. be (CF&PS3P(CF&, was nearly quantitative. It was The Thiophosphinic Acid Iodide. The cleavage of heated with more than one H2S per mole, in the same 2.76 mmoles of (CF&PSP(CF3), by 2.76 mmoles of IZ began slowly in the dark at 25", but went very rapidly ( 5 ) R. C. Dobbie, L. F. Doty, and R. G . Cavell, J. A m . Chem. Soc., 90, 201 5 (1 968), kindly shown to us by Dr. Cavell prior to publication. in sunlight (through a Pyrex wall), leading to almost (6) R. G.Cavell and H. J. Emelbus, J . Chem. SOC.,5896 (1964). complete disappearance of the iodine color. The (7) F. W. Bennett, H. J. Emelbus, and R. N. Haszeldine, ibid., 3896 product (CF&PI (2.55 mmoles; 92.4% yield) passed (1954). amount, this recycle of the diphosphine disulfide would imply an infinite series summing up to a 64% yield of (CF3)2PSP(CF3)L;actually, the yield was 66.5 % Most probably, another group exchange to form P2(CF& and (CF3)?PS4P(CF& was not significant, but the results could have been affected in some small degree by entrapment of P-I material in the nonvolatile products, or by some formation of Hg-S-P compounds. The acid chloride (CF&P(S)Cl was made quite easily by the action of chlorine on (CF&PSP(CF&, by heating (CF&PC13 with silver sulfide, or by the Al2Cl6-cata1yzed direct addition of sulfur to (CF&PCl at 100". The direct addition of sulfur also converted CF3PClz to CF8P(S)C12,again with A12C16. The thioester (CF&P(S)SCH3 could be made from CHBSH and (CF&P(S)Cl by using trimethylamine to remove HCI, but, when the acid (CF&PS2H was sought by the analogous reaction of H2S, the product of interest seemed to be the salt (CH3)3NH+(CF3)2PS2-, from which the very strong acid (CF&PSZH could not be liberated by hydrogen chloride. The acid amides were made easily by the ammonolysis or aminolysis of the appropriate chlorides. Our attempts to make them by the direct addition of sulfur to the aminophosphines did not succeed at 150": (CF&PN(CH& failed to react, while (CF3)?PNH2 (inert at lower temperatures) went primarily to nonvolatile products. However, we have been informed of experiments in which these aminophosphines added sulfur at 150 and 180°, respectively, to form the desired acid amidese5 The discrepancy most probably relates to the presence or absence of catalytic impurities, such as ammonia. Indeed, our observation that (CF3)?P(S)NH2 liberated HCF3 on standing (contrary to ref 5) would be explained if our sample contained a faint trace of ammonia, less likely to be present in the product of the sulfur-addition process. Ammonia would act like NaOH in basic hydrolysis and be regenerated as the compound is destroyed. This seems more likely than any base action by the NH, group, for its poor base action is proved by the slow reaction of (CF3)2P(S)NH2with high-pressure HCI. The basic hydrolysis of our (CF&P(S)X compounds gave only one HCFa per mole, and CF3P(S)X2 gave none. Thus we confirm that such basic hydrolyses stop at the phosphonate stage.6 Furthermore, the P=S bond also resists bases, for the basic solutions failed to liberate H2S upon acidification. However, the strong-acid hydrolysis of (CF&P(S)X compounds at 140" did produce HzS and one HCF3, meaning that hydrolysis under such conditions went to the CF3PO(OH)zstage and no farther.'

Journal of the American Chemical Society / 90:8 / April I O , 1968

201 3

through a high-vacuum trap at -60", where the desired product (CF3)zP(S)I (2.65 mmoles; 96% yield) was condensed, The slight iodine color of this product was removed by zinc dust at 25". The compound melted at 30", with decomposition affecting the observation. On standing at 20", the thiophosphinic iodide slowly liquefied with liberation of iodine, with a decrease of rate suggesting an approach toward equilibrium. Accordingly, the mixture was shaken with mercury for 1 hr at 25". The volatile product of 0.382 mmole of (CF3)2P(S)I was 0.127 mmole of (CF3)2PSP(CFs)z, representing 66.5 % of the original (CF3)2Punits. The character of the nonvolatile material was not studied. The action of H2S (in excess) upon (CF&P(S)I occurred readily below room temperature and led to a fraction somewhat less volatile and far lower melting than the pure (CF3)zPSzHwhich was obtained later by another method. However, the formation of this substance in major yield was proved by the characteristic infrared peaks-accurately the same as for the pure acid but showing also some impurities. It was not very difficult to eliminate the (CF,),PSH byproduct, but much harder to remove the condensation product of this, namely (CF3)2PSP(CF3)z. The Thiophosphinic Acid Amide. A 0.74-mmole sample of (CF&P(S)Cl was added slowly through a stopcock to 2.48 mmoles of ammonia in an 1100-ml bulb, immediately precipitating ammonium chloride. The reflux column was used at -35" to remove the excess ammonia and a trace of HCFs; then at -25" it delivered 0.425 mmole (57% yield) of apparently pure (CF3)aP(S)NH2. The nonvolatile liquid remaining in the bulb could well have been mostly the condensation product [(CF3)zPS]zNH, but it was not investigated further. The Thiophosphinic Acid Dimethylamide. The reaction of 2.08 mmoles of (CF3)2P(S)Clwith 4.17 mmoles of (CH3)zNH occurred in a sealed tube during a slow warming toward 25", but, in order to ensure participation by occluded reactants, the tube was heated for 10 hr at 60". High-vacuum fractional condensation delivered 0.067 mmole of HCF3 2nd 0.041 mmole of unused (CF&P(S)Cl. The yield of pure (CF&P(S)N(CH3),, delivered from the high-vacuum reflux column at -25", was 1.81 mmoles, representing 89% of the consumed (CF3),P(S)C1. The HC1 cleavage of this amide represented an elementary analysis, as shown by the following equation with millimole stoichiometry. (CFM'o)N(CHdz 0.173 -0.053 0.120

+ 2HC1+ 3.70 -3.46

(CF3)zP(S)CI 0.116

+ (CHahNHzCI

0.24

This experiment was carried on in a tube so small as to develop 11 atm pressure of HC1, and went for 3 days at 100". Apparently the conditions for complete cleavage of the amide would be far more stringent. The Methyl Dithioester. A gaseous mixture of 2.03 mmoles each of CH3SH and (CH&N was slowly introduced into 2.03 mmoles of (CF3)2P(S)C1, contained in a 1000-ml bulb at 25". The precipitation of (CH3)3NHCl was immediate. The volatile product was purified by the high-vacuum reflux column at -20":

1.70 mmoles (83.7 % yield) of (CF3)2P(S)SCH3. This product failed to react with HzS at 9 atm and 110". The Thiophosphonic Acid Dichloride. A sealed tube containing 6.35 mmoles of CF3PC12, 6.25 mg-atoms of sulfur, and 0.20 mmole of vacuum-sublimed AlZCle was heated to 100" to initiate a reaction which then proceeded vigorously. The reflux column was operated at -30" (to avoid formation of solids), delivering 5.90 mmoles of CF3P(S)C12(93 % yield). Table I. Confirmation of Molecular Formulas -Mol wt-. Obsd Calcd

Formula (CFI)ZPSZH (CFa)zPSzCH3

HCF3 per mole

S per mole

CI per

mole

2 2 7 ~229 247 248

0.991* 0.963' 0.985HzSc 0.961CH3SH (CF3)zP(S)Cl 237 236.5 0.995b 0.96d 1 .007d (CF3)zPCh 275 275.5 0.994 ... (CF3)zP(S)NH* 221 217 1.002d 0.97d (CF~)ZP(S)N(CH~)Z 248 245 l . O l b 0.95' CF3P(S)CIz 202 203 None 1.006c CF3PS[N(CH3)]z 226 220 None 0 . 9 9 7 a Equiv wt by titration, 230. * Basic hydrolysis. 140". By CF3COOH at 140".

...

... ...

1.017 3.000

...

...

2.00

By 6 M HCI at

Table 11. Volatility Constants

Formula

Nernst constants A B C

(CF3)zPSzH 2184 (CFs)zP(S)SCH3 2435 1849 (CF~)ZP(S)C~ (CF~)ZP(S)N(CH~)~ 2658 2026 CFaP(S)Clz CF3P(S)[N(CHa)z]z 3278

0.00500 0.00530 0.00500 0,00593 0.00550 0.00580

6.0330 6.4564 5.6678 7.1046 5.5578 7.7610

Trouton Bp, constant, "C eu 105.3 133.6 61.0 147.2 85.7 211.0

21.2 21.1 21.3 20.8 20.6 22.7

Table 111. Volatility Data (CF&PS2H, solid: log P = 8.903 - 2188/T; 7.81 mm at O.O", 14.94 mm at 10.0", 17.48 mm at 12.5", and 19.20 mm at 14.0"; calcd, 7.81, 14.98, 17.51, and 19.20; calcd mp 18'. Liquid: Temp, "C 18.0 23.9 27.4 32.0 35.6 41.3 P o b & Inm 24.40 33.30 39.82 49.80 59.79 76.90 Poslod,mm 24.44 33.33 39.80 49.89 59.77 76.91 (CF3)zP(S)SCHa: Temp, "C 0.00 Pobsd,mm 2.28 Poaiod,mm 2.28

11.4 4.85 4.85

26.0 11.6 11.6

46.9 34.2 34.3

70.1 95.5 95.5

(CF3)zP(S)C1,solid: logP = 8.778 - 1891/T; 3.16mmat -44.7', 6.0 mm at -37.0", and 17.6 mm at -22.1"; calcd 3.16, 5.9, Liquid: and 17.6; calcd mp -21.4". -12.0 -5.4 0.05 6.4 11.9 Temp, "C -20.0 Pabsd, mm 20.2 32.5 47.0 62.5 86.6 112.9 112.8 32.5 47.0 62.6 86.6 P o a l c d , mm 20.2 (CF~)ZP(S)N(CH~)Z : Temp, "C 0.00 Pabsd, mm 1.04 P o a l o d , mm 1.04 CF,P(S)CIz: Temp,"C Pobsd, mm Pcalod, mm

9.5 2.07 2.10

-21.2 6.37 6.37

-11.0 12.37 12.37

CF3PS[N(CHa)zlz: Temp, "C 40.6 Pobsd,mm 0.73 Peaiod.mm 0.73

51.6 1.51 1.51

Gosling, Burg

20.5 4.21 4.24 0.00 23.7 23.7 62.7 3.01 2.98

32.7 55.2 8.92 29.7 8.92 29.7 6.1 33.0 33.0

19.4 64.7 64.7

63.3 42.8 42.8 40.3 162.0 162.1

70.6 4.68 4.68

Bis(trij9uorornethyl)dithiophosphinic Acid

2014 Table IV. Infrared Spectra of Seven SPv Compounds Expected mode

S (CF&PSH

S (CF3)zPSCHs

S (CFa)zPCl

N-H, C-H, or S-H, v

2681 (1.2) 2566 (0.5)

2944 ( 1.5) 2845?

...

3515 (4.1) 3407 (4.4)

...

1438 (0.8)

...

1538 (6.1)

CH3 or "z,6 C-F, v C-N, v CHI, P P-x, v "2, P S-H bend CH, wag P=S, v CF3, 6-e S-CHs, v CF3, 6-a P-CF,, v P=S bend CF3, P CF3, wag

1207 (52) 1174 (78)

... 524